Chin. J . Chrm. Eng.

,15(3) 433-438 (2007)

The Morphology, Phase Composition and Effect of Corrosion Product

on Simulated Archaeological Iron*

WANG Zise(€ %&)", XU Chunchun(iFZ$ CAO Xis(% b)"and XU Ben($$?@)"

a School of Material Science and Engineering, Beijing University of Chemical Technology, Beijing 100029,China
The State Key Laboratory for Metallic Corrosion and Protection, Shenyang 110015,China

Abstract The immersion corrosion of archaeological iron in solution (0.06mol.L- I NaC1+0.03mol.L I Na2S04+
~

0.0 1mol.L-' NaHC03) simulating soil water composition was presented. The evolution of archaeological iron from
iron to iron oxide and to iron oxy-hydroxides compounds was investigated by scanning electron microscope (SEM)
and X-ray diffraction (XRD) analysis. According to the morphology, phase composition, and transformation proc-
ess, the contributions of each corrosion product to archaeological iron were discussed.
Keywords archaeological iron, corrosion, scanning electron microscope, X-ray diffraction

1 INTRODUCTION iron by analyzing the transforming process.

The deterioration of archaeological iron due to
corrosion is a well-known problem. -k variety of cor-
rosion products can be formed depending on the spe-
cies present in their environment (soil, sea water, at-
mosphere, etc.). To find out the evolution of iron rust
is a real challenge since it is difficult to clearly iden-
tify the contributions of each corrosion product[ 11.
Although the evolution of iron oxides and
oxy-hydroxides in different aeration conditions and
pH have already been tested[2], their effects on the
stability of archaeological iron have not been system-
atically studied. In addition, at present the most seri-
ous problem during restoration of iron artifacts is to
find a suitable chemical treatment for their preserva-
tion[31. Thus, the elucidation of the phase composition
of the corrosion layers appears to be very important
from the viewpoint of choosing proper chemical re-
agents to treat archaeological iron objects for their
preservation.
The main phases constituting the rust layers
formed on iron antiquities exposed to soil corrosion
are rnagnetite(Fe304), goethite(a-FeOOH), lepido-
crocite(y-FeOOH), and akaganeite (0-FeOOH). Gen-
erally, iron rusts can be classified into harmful and
harmless ones according to the effects on the corro-
sion of iron[4,5]. Harmless rusts usually do not de-
velop on the iron matrix in the normal environment
and do not need to be removed from the iron antiqui-
ties. Harmful rusts, on the contrary, are able to accel-
erate the further corrosion of the metal core i n the
normal environment and considered to be the main
trouble for the iron relics.
I n this article, scanning electron microscope
(SEM) and X-ray diffraction (XRD) will be utilized to
investigate the morphology and structure of corrosion
product on archaeological iron. Aqueous solution
whose composition is mostly the same as that in Chi-
nese soil will be adopted. An attempt was made to
distinguish the effects of different kinds of corrosion
products on the corrosion evolution of archaeological
2 EXPERIMENTAL
Cast iron, with the mass composition (96)C 3.00,
Si 1.84, S 0.089, P 0.098, Mn 0.82, Fe balance, was
used to simulate archaeological iron. The samples
were processed to be 15mmX 15mmX3mm coupon.
The surfaces of these samples were polished with em-
ery papers up to 1000 grade followed by thorough
rinsing in acetone and de-ionized water.
C1 +
I
The corrosion media were 0.06mol.L
0.03m0l.L~' SO:- +O.O 1mo1.L 'HCO; solutions
diluted from AR grade NaCl, Na2S04, NaHCO? with
de-ionized water, H2S04 and NaOH having 20%
volume ratio were used to adjust pH to 7 in order to
eliminate the effects of pH on the experiment.
The solution was de-oxygenated with purified ni-
trogen for 30min previously to each immersion to
simulate the state of soil that lacked oxygen in deep
earth. The specimens were fixed on the brackets at the
bottom of beaker and the beaker was made to be air-
proof. After immersion for a certain period, the
specimens were dried in nitrogen for SEM observation
and XRD test. X-ray diffraction (Model ?MAX) was
set up as follows: scan speed lO(").min , from 3" to
90", with Cu K, as its radiation source operated at
50kV and 200mA.
3 RESULTS AND DISSCUSION
3.1 Corrosion products on the archaeological iron
after immersion for 48 days
Figure 1 showed the images of corrosion products
on the upper surface of the specimen after immersion
for 48 days. It seems that corrosion products are dis-
tributed loosely on the upper surface of the specimen
with grain-like shape, and the grain diameter are in the
order of 1. e l Spm. Because of the loose distribution
of corrosion product, a lot of cavities or gaps are left,
even iron matrix is exposed on the local surface.

Received 2006-06-07, accepted 2007-01 -15.

* Supported by the National Key Technologies R&D Program of the 10th Five-Year Plan Period (No.2001BA805B01).
** To whom correspondence should be addressed. E-mail: chunchunxu@263.net
434 Chin. J. Ch. E. (Vol. 15, No.3)

Figure I SEM image of corrosion products on the upper

surfirce of the specimen after immersion for 48 days

The composition of surface corrosion products

was identified by means of X-ray diffraction; the re-
sult was shown in Fig.2. Obviously, the spectra are
mainly of goethite and some iron matrix. Accordingly,
the grain-like corrosion product formed on the upper
surface of the specimen after immersion for 48 days
mainly consisted of goethite. But goethite didn't cover
the iron surface; its loose distribution can be proved
by the appearance of spectra of iron matrix. Here the Figure 3 SEM image of corrosion products on the under-
composition of grain-like substance, not confirmed in surface of the specimen after immersion for 48 days
Dillmann's study on atmospheric corrosion of iron[6],
proved to be goethite. Zhu et a1.[7] observed that
goethite was stalactite-like on the surface of iron arti-
facts. Possibly, the difference is because of the shorter
immersion period (48d) in this experiment than the
buried period (several thousand years) of genuine iron
artifacts in the soil. The grain of goethite was tiny
since it formed in the initial corrosion stages on the
iron surface.
I I I

20 40 60 XO
20, ( )
"

Fc: iron matrix

VI
5 800 G: goethite Figure 4 XRD spectrum of corrosion product on the
undersurface of the specimen after immersion for 48 days

grain was coarse to naked eyes. After the outer layer

I t I I I I I I
20 30 40 50 60 70 80
28, ('I ) dark green. The layers can be easily peeled off from
Figure 2 XRD spectrum of corrosion products on the up-
per surface of the specimen after immersion for 48 days
Figure 3 showed the images of corrosion product
on the undersurface of the specimen after immersion
for 48 days. The corrosion product was tiny scale-like
shape with loose distribution and some petal-like
morphology was also present in partial corrosion
product. Fig.4 revealed the result of XRD test to this
scale-like rust; it was proved to be primarily lepido-
crocite, which coincides with author's former study[8]
on the corrosion products of archaeological iron.
3.2 Corrosion products on the simulated archaeo-
logical iron after immersion for 138 days
After immersion for 138 days, the surface of
specimen was covered with a thick layer of corrosion
product, the outer one of which was brown, and the
of the rust removed by needle, there were two other
I
layers below. The inner one adhering to the metal core
90 was orange in color, while the middle one was little
each other.
Figure 5 showed the images of corrosion prod-
ucts on the undersurface of the specimen after immer-
sion for 138 days. It can be observed that the inner
layer is covered with lots of acicular substance or
square acicular substance; while the middle layer is
leprose, similar to Fig.3(a); the outer layer also shows
leprose shape, but the scales have become thicker than
that found in the inner layer. Moreover, from the mag-
nified image Fig.S(d) some scales covered with
granular substance are observed and the corrosion
product becomes stalactitic.
From the XRD analysis on the thick layer of
corrosion product in Fig.6, it seems that there are sev-
eral diffraction peaks of goethite, a few diffraction
peaks of lepidocrocite, and some separate diffraction
peaks of hematite or magnetite. It was confirmed ear-
lier in this article that the stalactite-like corrosion
product was goethite (Fig.2), and the scale-like one

June, 2007
 The Morphology, Phase Composition and Effect of Corrosion Product on Simulated Archaeological Iron 435

middle layer on the undersurface of the specimen after

immersion for 138 days consists of lepidocrocite, and
stalactitic outer layer consists of goethite. But in Fig.6
the diffraction peaks of lepidocrocite are less than that
of goethite; this is probably because X-ray hardly
penetrates through the thick and compact stalactitic
goethite into the middle layer of lepidocrocite result-
ing in the few diffraction peaks of lepidocrocite. In
addition, the corrosion products on the upper surface
of the specimen after immersion for 138 days were
consisted mainly of goethite, as was shown in Fig.7.

G
I6Oc G 1 G: eoethitr
c
vi

=3 120
s
.z 80
E
c
.E 40

0
20 30 40 50 60 70 80 00
20, )
('I

Figure 7 XFtD spectrum of corrosion product on the up-

per surface of the specimen after immersion for 138 days

According to the previous researches on the cor-

rosion of iron objects[91, electrochemical process oc-
curs on the iron in the environment; the formed rusts
are in fact a sort of precipitation with very fine rust
grain and loose distribution. As a result, the oxygen
and water are easy to approach the metal core and ag-
gravate the corrosion.
Fe 4Fez++ 2e (1)
I
-0, + H,O + 2e-20H- (2)
2 -
Fez' + 20H- dFe(OH), (3)
(d) The magnified image of the small frame on (c)
Figure 5 SEM images of corrosion products on the un- Fe(OH), + O2+Fe,O, . H,O or 2FeOOH
(4)
dersurface of the specimen after immersion for 138 days
In the humid environment iron is corroded to
y-FeOOH in the beginning, then with the interaction
350 of water and oxygen y-FeOOH partly or entirely
<;: goethite transforms into stable a-FeOOH. This coincides with
B: hematite
e 250 C: magnetite the experimental result. The specimen was parallel to
,g 200
the bottom of beaker in the solution, so the diffusion
of oxygen from the mouth of the beaker to the upper
s
I
II

.-
150
surface of specimen was easier than its diffusion to the
I00 undersurface. Therefore, after immersion for a long
50 time, the undersurface of specimen was mainly cov-
0 '
10
1

20
I

30
I

40
I

50 60
I
70
ered with unstable rust y-FeOOH, and later it was
transformed into more stable goethite due to the
2H,(") abundant supply of oxygen; thus, the rust on the upper
Figure 6 XRD spectrum of corrosion product on the un- surface was mainly goethite. Attributed to the diffu-
dersurface of the specimen after immersion for 138 days sion of oxygen, the leprose y-FeOOH, and a-FeOOH
transforming to stalactitic phase were observed on the
undersurface of specimen immersed for 138 days.
was lepidocrocite (Fig.4). Consequently, together with However, what on earth were the acicular substance
the SEM images in Fig.5 and the above-mentioned and square acicular substance in the inner layer?
X R D spectrum, it can be concluded that the leprose After dried in nitrogen, it was found that the
Chin. J. Ch. E. 15(3) 433 (2007)
436
corrosion product on the specimen immersed for 138
days cracked and became easy to be peeled off. XRD
was conducted onto the inner side of the corrosion
product, specifically the side clinging to the iron ma-
trix. The result was shown in Fig.8, and the micro-
structure shown in Fig.S(a).
A: akaganeite
G: goethite
C:magnetite
10 20 30 40 50 60 70
20, ( )
Figure 8 XRD spectrum of corrosion product on the inner
layer of the specimen after immersion for 138 days
XRD analysis showed that the inner layer of the
product was mainly composed of a-FeOOH and a lit-
tle a-FeOOH and Fe304.This is because lepidocrocite
that formed on the iron surface not only react with the
oxygen and water and become goethite, but also can
act on Fez+dissolved from the iron and form magnetite.
At the same time, this reaction just expresses the
process of cathodic depolarization for the continual
eroding of iron artifacts with rusts after exhumation.
8y- FeOOH + Fezf2e-3Fe30, +4H20
Figure 5(a) showed the morphology of corrosion
product that was closely clinging to the metal core. It
seems that the surface of iron is scattered with many
acicular substances; below them a lot of little cube
grains are distributed. Fe304has a cubic spinelle-like
structure. From the results of XRD test it can be
confirmed that the little cube grains distributed on the
iron surfxe are FeBO,. This is consistent with Matsui
et ul.'s observation[ 101 to the Fe304grains on the rusts
of real iron antiquities.
The dissolusion process of archaeological iron in
the media containing CIK can be expressed[ 111:
Fe+CI- ,
k,,
k ~ , 'FeCI,, +e
FeCI,,, >' >Fe:i + C1- + e (7)
When the specimen is dried, the humidity on the
surface of cast iron that is in dissolution state largely
declines; thus, the dense FeC12 liquor is concentrated
under the thick rusts or in the pitting cavities. This
leads to the phenomenon of "weeping" that shows
yellow droplets containing ferrous and ferric ions on
the surface of the artifacts[ 121. During burial, the po-
rous oxide layer is filled with an aqueous electrolyte
containing a large amount of ferrous ions. Hydrolysis
of ferrous ions lowers the pH, and chloride ions will in
most cases be the dominant counter ion due to its high
mobility and availability[ 131. Thus, the local high
concentration of chloride combined with a low pH is
June, 2007
Chin. J. Ch. E. (Vol. 15, No.3)
favorable for akaganeite formation and makes aka-
ganeite of special relevance to the stability of ar-
chaeological objects. As soon as the specimen con-
tacted with air, P-FeOOH will be produced between
rusts and iron because of the existence of C1- and oxi-
dation proceeding in acid conditions[ 141:
Fe2' + 0, + H,O CIK
,p - FeOOH + H+ (8)
The reason for the presence of P-FeOOH on the
XRD spectrum of corrosion product on the undersur-
face of the specimen corroded for 48 days is also ex-
plained here.
Briefly, in the humid environment, an active rust
y-FeOOH was formed on the archaeological cast iron
in the beginning; then with the action of water and
oxygen, y-FeOOH was transformed into a-FeOOH,
80 which is more stable than y-FeOOH. At the same time,
y-FeOOH that was formed on the matrix of cast iron
reacted with Fe2+ leads to the formation of Fe304.
Owing to drying, FeC1, were concentrated on the
matrix of samples so that Fez+ was oxided into
P-FeOOH with the action of C1-. As a result, the cor-
rosion product of cast iron in the solution for I38 days
was made up of three layers: a-FeOOH, Fe304,and a
little P-FeOOH in the inner layer; y-FeOOH in the
middle layer; and a-FeOOH in the outer layer.
3.3 Effects of rusts on the archaeological iron
3.3.1 Goethite (a-Fe00H)
Production free energy of goethite at 25% is
(5) -495.748kJ-mol-', which is less than that of lepido-
crocite -470.25kJ.mol-I[ 151. This is to say, goethite
is relatively stabler than lepidocrocite in thermody-
namics. Besides, from the SEM images [Fig.S(d)] it
can be seen that the layer of goethite, with the crust
like stalactite, has a good continuity and compactness
in texture. So it is assumed to be a non-reactive phase
with an important protection[ 16) owing to its ability to
effectively prevent the iron matrix suffering from at-
tacks of other environmental factors such as oxygen,
moisture as well as polluted gases.
3.3.2 Lepidocrocite (y-FeOOH)
Lepidocrocite has a loose structure with compli-
cated leprose or petallike shape (Fig.3). It is porous
and similar in absorbency to the skeleton of sponge.
(6) Water, electrolyte, oxygen, or other polluted gas in the
environment can be absorbed and reserved. Air pol-
lutants like SO2and NO2 dissolve in the fluid film on
the surface of iron objects resulting in the decrease of
pH and the acceleration of y-FeOOH dissolution.
Therefore, the concentration of Fe3+will be increased:
y - FeOOH + 3H+ dFe3++ 2 H 2 0 (9)
Being oxidant, Fe3+and its ramification, whose
oxidation ability increases with concentration of H+,
can be the catalyzer and activator to the corrosion of
iron core. Fe3+is able to either directly oxide the iron
to be Fe2+,or induce oxygen acting on Fe2+andaccel-
erate the transformation from Fe" to Fe3+.As long as
the medium and environment are propitious to the
dissolution of lepidocrocite and production of Fe3+,
 The Morphology, Phase Composition and Effect of Corrosion Product on Simulated Archaeological Iron
the process mentioned previously will be cycled, and
the continual corrosion of iron will be induced below
the rusts[ 171. Consequently, lepidocrocite is a kind of
harmful rusts to the preservation of iron antiquities. It
is strongly suggested that the lepidocrocite should be
removed from iron antiquities immediately after ex-
cavated, even compulsive desiccation or chemical
disposal for stabilization can be used to restrain its
development.
3.3.3 Akaganeite (ljl-FeOOH)
Formation of ferric oxy-hydroxides is associated
with volume expansion, which cause cracks and exfo-
liation of the corrosion layers or even the whole arti-
fact, which in turn may accelerate the deterioration. At
the position of fracture, an orange powder can be ob-
served, which has been identified as akaganeite on
several artifacts[ 181. Akaganeite, with the molecular
formula Fe00.833(OH)I.167C10.167[ 131, is mainly acic-
ular or virgulate in appearance. Its structure is even
not as compact as lepidocrocite, and the interspaces
among the acicular are relatively large. The absorption
or storage ability is stronger than lepidocrocite. As a
result, different kinds of polluted gas, oxygen or
moisture are easy to be reserved. This supplies favor-
able conditions for the occurrence of local corrosion to
a great extent[ 191.
As is known, C1- was considered to be one of
the most important factors to the preservation of ar-
chaeological iron[20], which not only involve the
chloride that was formed in the environment (sea wa-
ter, soil) before exhumation, but also the chloride ions
trapped in the lattice of akaganeite[21]. It is obvious
that there is chloride element in the molecular struc-
ture of P-FeOOH, and the mass ratio comes to 6.24%.
When the iron antiquities are conserved, the dissolu-
tion of polluted acid gas will make the solution acidic,
thus induce akaganeite to dissolve, and chloride ele-
ment will be released in the form of ions. Simultane-
ously, the phase transformation like decompounding
-
due to drying can also bring on the release of chloride
element[ 131:
2Fe00.833(0H),.,67 c10.167
Fe,O, + 0.333HC1+ H,O (10)
the released chloride dissolves in the acid fluid film,
together with H+ and the oxygen which diffuses into
the rust, forms a condition capable of accelerating the
corrosion of iron.
The existence of akaganeite is a risk for ironwork,
and chloride element carried by akaganeite is the main
potential trouble. So akaganeite is another kind of
harmful rusts to ironware and must be stabilized or
removed in the preservation. At present, conservators
have been using various treatments for removal of
chlorides from artifacts including different desalina-
tion and heating treatments. Dechlorination is used to
decrease or remove chloride ions by washing and im-
mersion in alkaline solution or by electrochemical
technologies.
3.3.4 Magnetite (Fe304)and hema;tf:te(Fe203)
Lepidocrocite reacted with Fe which dissolved
from iron matrix to be magnetite, or become hematite
437
in certain conditions, for example, desiccation. This is
also the reason Fe203 and Fe304 exist in the XRD
spectrum of corrosion product on the undersurface of
the specimen after immersion for 138 days (Fig.6).
Moreover, in dry atmosphere and dry soil, rusts
formed on the iron objects are mainly oxides. As an-
other kind of rust on ironwork, iron oxide is stabler
than iron oxy-hydroxides. Fez03 and Fe304have the
production free energy - 822.16kJ.mol - 1 and
- 1117.13kJ~mol-’, respectively, which are far less
than that of y-FeOOH and a-FeOOH. They also have
very high lattice binding energy. Consequently, Fe203
and Fe304areharmless rusts; they are able to prevent
iron artifacts from farther corrosion.
Iron objects usually need to be compulsively
dried immediately after exhumation. The purpose for
this is transforming the unstable rusts to stable rusts
like Fe203 and Fe304 to avoid the further development
of harmful rust. Besides, at present there are many
chemical stabilization methods for the preservation of
iron antiquities. However, no matter what kinds of
methods, they are all intended to accelerate the trans-
formation of unstable, active and loose rusts into sta-
ble and compact rusts[22] so that the protection can be
arrived.
4 CONCLUSIONS
(1) Corrosion products on the upper surface of
the specimen after immersion for 48 days were mainly
a-FeOOH, which was grain-like in appearance with
loose distribution; while corrosion products on the
under surface of the specimen after immersion for 48
days were mainly y-FeOOH, with leprose or petallike
shape.
(2) Corrosion product of archaeological iron in
the solution for 138 days was made up of three layers:
a-FeOOH, Fe304 and a little P-FeOOH in the inner
layer; y-FeOOH in the middle layer; and a-FeOOH in
the outer layer.
(3) In the rust of iron antiquities, the layer of
a-FeOOH, with the shape like stalactite, is able to
prevent the iron matrix suffering from attacks of other
environmental factors because of its good continuity
and compactness. y-FeOOH has a complicated leprose
or petal-like structure; it can decompose to Fe3+under
acidic condition and able to accelerate the corrosion of
iron matrix. The molecular structure of P-FeOOH in-
cludes chloride element which, if released, will induce
the deterioration of iron core. So y-FeOOH and
P-FeOOH are harmful rusts to the preservation of iron
antiquities. Moreover, iron oxide like Fe203 and Fe304,
which are more stable than iron hydroxide, is harmless
to iron objects.
REFERENCES
1 Peulon, S., Antony, H., Legrand, L., Chausse, A., “Thin
layers of iron corrosion products electrochemically de-
posited on inert substrates: Synthesis and behaviour”,
Electrochim. Acta, 49, 2891-2899(2004).
2 Neff, D., Dillmann, P., Gurlet, L.B., Beranger, G., “Cor-
rosion of iron archaeological artefacts in soil: Charac-
terisation of the corrosion system”, Corros. Sci., 47, 5 I5-
535(2005).
438 Chin. J. Ch. E. (Vol. 15, No.3)
3 Novakova, A.A., Gendler, T.S., Manyurova, N.D., Tur-
ishcheva, R.A., “A Mossbauer spectroscopy study of the
corrosion products formed at an iron surface in soil’’,
Corros. Sci., 39(9),1585-1594(1997).
4 Huang, Y.L., Lin, B.X., Wang, C.S., Li, J.H., Minoru, S.,
“The structural characteristics of corrosion products on the
ancient ironware”, Sci. Conserv. Archaeol., 8( l), 26-27
(1996). (in Chinese)
5 Liao, Y., “ Research on erosion mechanism of bronze
cultural relics and transform agent on harmful rust”, Sci.
Conserv. Archaeol., 15(2),21-23(2003). (in Chinese)
6 Dillmann, Ph., Mazaudier, F., Hoerl,S., “Advances in
understanding atmospheric corrosion of iron. I. Rust
characterisation of ancient ferrous artefacts exposed to
indoor atmospheric corrosion”, Corros. Sci., 46, 1401-
1429(:2004).
7 Zhu, H.F., Zhou, H., Cai, L.K., “Study on chemical stabi-
lized conversion of rust layers in conservation of ar-
chaeological iron”, Sci. Conserv. Archaeol., 7(2), 1-
I l(1995). (in Chinese)
8 Wang, Z.S., Xu, C.C., Wang, J.L., “Appearance and for-
mation process of corrosion production archaeological
iron in simulated soil media”, J. Chem. Ind. Eng. (China),
56( 12), 2373-2379(2005). (in Chinese)
9 Yu, F.Z., Corrosion Resistance of Metal Materials, Sci-
ence Press, Beijing (1982).(in Chinese)
10 Matsui, T., “Examination and conservation of ancient
iron”, In: International Congress on the Conservation and
Preservation for Archaeological Objects, Japan (2002).
11 Kim, .J.D., Pyun, S.I., “The effects of applied potential
and chloride ion on the repassivation kinetics of pure
iron”, Corros. Sci., 38, 1093-1 102(1996).
12 Turgoose, S., “Post-excavation changes in iron antiqui-
ties”, Stud. Consent., 27,97-101( 1982).
June, 2007
13 Kenny, S.A., Kurt, N., Jiang, J.Z., Bente, L., Jonathan,
C.H., Poul, N., Jettie, V.L., “On the akagankite crystal
structure, phase transformations and possible role in
post-excavational corrosion of iron artifacts”, Corros.
Sci., 45,2563-2575(2003).
14 Zhang, H.J., Zhang, X.L., “Research on the principle of
transform and protect of ferric cultural relics”, Cultural
Relics and Archaeology of Inner Mongolia, 2, 90-
92(2003). (in Chinese)
15 Zhu, H.F., Zhou, H., Cai, L.K., “Corrosion inhibition of
silicate in the desalination cleaning solution for iron an-
tiquities”, Sci. Conserv. Archaeol., 14 (Suppl.), 52-
62(2002).(in Chinese)
16 Yamashita, M., Miyuki, H., Matsuda, Y., Nagano, H.,
Misawa, T., “The long term growth of the protective rust
layer formed on weathering steel by atmospheric corrosion
during a quarter of a century”, Corros. Sci., 36, 283-
299( 1994).
17 Zhang, A.F., “Effects on atmospheric corrosion of iron”,
Mater. Protec., 22(2), 16-19( 1989).(in Chinese)
18 North, N.A., “Corrosion products on marine iron”, Stud.
Conserv., 27,75-83( 1982).
19 Tan, W., Xu, B.S., Han, W.Z., Zhong,Q.P., “Analysis on
rusts of carbon steel subjected to 2+30 years exposure
in South-East coastal atmosphere”, Corrosion Science
and Protection Technology, 15(2),113-116(2003). (in
Chinese)
20 Turgoose, S., “The corrosion of archaeological iron during
burial and treatment”, Stud. Conserv., 30, 13-18( 1985).
21 David, AS., Nigel, J.S., “The washing of fragile iron ar-
tifacts”, Stud. Consent., 32, 73-76( 1987).
22 Yu, K., Xu, C.C., “Inhibitive effect of sodium molybdate
on iron relics”, J. Beijing Univ. Chem. Tech. (Natural
Science Edition), 31(4), 41-44(2004). (in Chinese)